Chapter 5 Shape Selective Recognition of T•A Base Pairs by Hairpin Polyamides Containing Novel N-Terminal Pairings The text of Chapter 5A was taken in part from a manuscript coauthored with Michael Marques, Ray Doss, and Professor Peter Dervan (California Institute of Technology). Foister, S.; Marques, M. A.; Doss, R. M.; Dervan, P. B. “Shape Selective Recognition of T·A Base Pairs by Hairpin Polyamides Containing N-Terminal 3- Methoxy (and 3-Chloro) Thiophene Residues.” Bioorg. Med. Chem., in press. The research efforts described in Chapter 5B are part of ongoing research efforts in the Dervan group. Work concerning novel N-terminal residues done in collaboration with Ray Doss and Michael Marques and efforts to characterize nuclear localization of novel residues were in collaboration with Ben Edelson and Tim Best. Application of novel residues to controlling expression of the hMDR gene is an ongoing collaboration with Chris Martin (Waring group, Cambridge). Application of novel residues to centromere staining is an ongoing collaboration with B. Edelson, T. Best (Dervan Group, Caltech) and Edward Ramos (Trask group, Washington).
28
Embed
Chapter 5 Shape Selective Recognition of T A Base Pairs by ...thesis.library.caltech.edu/1720/7/(07)ch5a_Title_text.pdfChapter 5 Shape Selective Recognition of T•A Base Pairs by
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Chapter 5
Shape Selective Recognition of T•A Base Pairs by Hairpin Polyamides Containing Novel N-Terminal Pairings
The text of Chapter 5A was taken in part from a manuscript coauthored with Michael Marques, Ray Doss, and Professor Peter Dervan (California Institute of Technology). Foister, S.; Marques, M. A.; Doss, R. M.; Dervan, P. B. “Shape Selective Recognition of T·A Base Pairs by Hairpin Polyamides Containing N-Terminal 3-Methoxy (and 3-Chloro) Thiophene Residues.” Bioorg. Med. Chem., in press. The research efforts described in Chapter 5B are part of ongoing research efforts in the Dervan group. Work concerning novel N-terminal residues done in collaboration with Ray Doss and Michael Marques and efforts to characterize nuclear localization of novel residues were in collaboration with Ben Edelson and Tim Best. Application of novel residues to controlling expression of the hMDR gene is an ongoing collaboration with Chris Martin (Waring group, Cambridge). Application of novel residues to centromere staining is an ongoing collaboration with B. Edelson, T. Best (Dervan Group, Caltech) and Edward Ramos (Trask group, Washington).
129
Abstract
The vast potential of polyamides as therapeutic agents or as tools for
genomic analysis is primarily due to their capacity to bind predetermined DNA
sequences with affinities and specificities rivaling naturally occurring proteins. The
pairing rules governing DNA recognition by minor groove-binding polyamides dictate
that internal cofacial pairings of Im and Hp with Py facilitate specific recognition of all
four Watson-Crick base pairs. The affinity and specificity of N-terminal Hp/Py
pairings, on the other hand, are substantially diminished relative to internal contexts.
Previous research efforts in the Dervan group examined the efficacy of a
series of 6-substituted-2-hydroxybenzamide residues as N-terminal replacements for
Hp. Substitution of the hydroxybenzamide scaffold was envisioned as a steric
means of constraining the rotational freedom of N-terminal residues, forcing the
hydroxyl recognition element into the floor of the minor groove. Modest T•A
selectivity at subnanomolar concentrations was demonstrated by a 6-methoxy-2-
hydroxybenzamide paired opposite Py; however, the synthesis of polyamides
containing this residue is at best challenging.
Collectively, the body of work on N-terminal recognition of the DNA minor
groove by polyamides suggests that the hydroxyl recognition element may not be
optimal for thymine discrimination by N-terminal pairings. Drawing upon the steric
component of thymine specificity shown by internal hydroxypyrrole and
methylthiophene heterocycles, a series of novel N-terminal residues was designed
to target T•A base pairs by a purely shape selective mechanism. Chapter 3A
describes the synthesis and characterization of these residues while Chapter 3B
130
discusses their influence on cellular localization and application in biological
collaborations.
131
Chapter 5A
Shape Selective Recognition of T·A Base Pairs by Hairpin Polyamides Containing N-Terminal 3-Methoxy (and 3-Chloro) Thiophene Residues
Background and Significance
Polyamides composed of N-methylpyrrole (Py), N-methylimidazole (Im), and
N-methylhydroxypyrrole (Hp) amino acids are crescent-shaped ligands that bind
sequence specifically in the minor groove of DNA and have the potential to modulate
gene expression. The specificity of DNA recognition arises from interactions
between the edges of the Watson-Crick base pairs and antiparallel aromatic amino
acid ring pairs oriented N→C with respect to the 5’→3’ direction of the DNA helix.1-3
Covalent head-to-tail linkage of two polyamide strands by γ-aminobutyric acid
constitutes the hairpin motif, in which opposing residues from each strand are locked
into cofacial pairs.4, 5 Im/Py distinguishes G•C from C•G and both of these from T•A
/ A•T base pairs while a Py/Py pair binds both T•A and A•T in preference to G•C /
C•G. The exocyclic amino group of guanine imparts G•C specificity to Im/Py pairs
through formation of a specific hydrogen bond with N3 of Im. Binding of Py/Py is
disfavored at G, C base pairs by destabilizing steric interactions between the C3-H
of Py and the guanine amino group.6, 7 The replacement of C3-H of one Py with
hydroxyl creates the Hp/Py pair which exploits the steric fit and hydrogen bond
acceptor potential of thymine-O2 as well as the destabilizing steric interaction with
the bulkier adenine ring to gain specificity for T•A.8,9
The above pairing rules have been used to design hundreds of synthetic
ligands that bind predetermined DNA sequences. However, many sequences
132
remain difficult to target, likely due to sequence dependent microstructure variations
in minor groove width or curvature. Furthermore, the specificity of cofacial aromatic
amino acid pairings depend on their context (position) within a given hairpin
polyamide. For example, Im/Py pairings show comparable specificity for G•C at both
terminal and internal positions.1 Conversely, Hp/Py pairings do not specify T•A at
the N-terminus of hairpin polyamides.10 The context dependence of Hp is
presumably a result of the conformational freedom inherent to an N-terminal
aromatic residue. The absence of a second “groove-anchoring” carboxamide allows
terminal rings to bind DNA in either of two conformations. For a terminal Hp residue,
a rotamer with the hydroxyl recognition element oriented away from the floor of the
minor groove could be stabilized by intramolecular hydrogen bonding between the
C3-OH and the carbonyl oxygen of the 2-carboxamide. For terminal 2-
hydroxybenzamide residues, some measure of T•A selectivity was recovered by
creating steric bulk at the 6-position to force the hydroxyl recognition element into
the groove.10 However, N-terminal pairings capable of binding T•A, with affinity and
specificity comparable to those of Im/Py for G•C, remain to be devised.
The fidelity of minor groove recognition by N-terminal Im/Py pairings in hairpin
polyamides can be rationalized by a combination of both stabilizing and destabilizing
forces which favors the rotamer with N3 in the groove and N-methyl out. Rotation of
a terminal Im residue in the opposite conformer, orienting N3 away from the minor
groove, would create unfavorable lone pair interactions with the proximal
carboxamide oxygen, disrupt a favorable hydrogen bond with the exocyclic amine of
G, and project an N-methyl group to the DNA floor which is presumably sterically
133
unfavorable. We address in this paper whether a T recognition element could be
designed using the asymmetric cleft of a T•A base pair as the basis for shape
selective discrimination. Recent work from our group has indicated that the
polarizable sulfur atom of thiophene heterocycles might serve this purpose.11
Our experimental design anticipated that substitution of the 3-position of a
thiophene-2-carboxamide scaffold could be used to favor an anti (“sulfur down”)
conformation at the N-terminus by disfavoring contact of the 3-substituent with the
floor of the minor groove (Figure 5.1). It was envisioned that the electronic
properties of the 3-substituent might be used to tune the polarization of the sulfur
atom, allowing a more complementary fit with thymine in the minor groove. We
attempt here to expand the repertoire of DNA sequences that can be targeted using
hairpin polyamides by investigating the DNA recognition properties of a series of N-
terminal residues consisting of 3-substituted-thiophene-2-carboxamide heterocycles.
Quantitative DNAse I footprinting was used to determine the affinity of eight novel N-
terminal 3-substituted thiophene rings residues, paired opposite Py, for each of the
four Watson-Crick base pairs (Figure 5.2). Ab initio computational modeling was
used to guide interpretation of the experimental results.
134
Figure 5.1 Proposed binding models for hairpin polyamides with 5’-TXTACA-3’ site. A circle enclosing two dots represents the N3 and O2 lone pairs of purines and pyrimidines, respectively. A circle containing an H represents the exocyclic amine of guanine. Putative hydrogen bonds are indicated by dashed lines. (A) Syn (‘sulfur down’) conformation of N-terminal residue. (B) Anti (‘sulfur up’) conformation of N-terminal residue.
135
Results
Monomer Synthesis (Figure 5.3)
Methyl 3-aminothiophene-2-carboxylate was Boc-protected and the resulting
ester was saponified to yield 3-[(tert-butoxy)carbonylamino]-2-thiophenecarboxylic
acid 11. Methyl 3-hydroxythiophene-2-carboxylate 12 was prepared by cyclization of
methylthioglycolate and methyl-2-chloroacrylate in methanolic sodium methoxide.12
Alkylation of 12 with iodomethane and subsequent hydrolysis of the methyl ester
gave 3-methoxy-2-thiophenecarboxylic acid 13. 3-Fluorothiophene-2-carboxylic acid
14 was synthesized as described previously.13 The remaining 3-substituted-
thiophene-2-carboxylic acids were obtained from commercial sources.
Figure 5.2 Experimental design for evaluation of novel N-terminal residues. (top) pCW15 plasmid design. (bottom) Ball and stick model of hairpin conjugates containing variable N-terminal residues. Shaded and non-shaded circles represent imidazole and pyrrole, respectively. A circle containing an S denotes a thiophene residue (1-8).
136
Polyamide Synthesis (Figure 5.4)
Polyamide resin R1 was prepared using manual solid phase synthetic
techniques described previously.14 Treatment of this resin with trifluoroacetic acid
(80% TFA in CH2Cl2) yielded a support-bound amine that was subsequently acylated
with the appropriate, HBTU-activated, thiophene-2-carboxylic acids. Acylation of R1
by 11 and removal of the Boc protecting group with TFA yielded resin R3 which was
cleaved with dimethylaminopropylamine (Dp) to give polyamide 3. Treatment of R3
with acetic anhydride prior to cleavage with Dp gave polyamide 4. The remaining
polyamides 1, 2, and 6-8 were cleaved from resin with Dp immediately following
acylation of the carboxylic acid. Treatment of 6 with sodium thiophenoxide in DMF
Figure 5.3 Synthesis of N-terminal thiophene building blocks. A) Synthesis of 3-[(tert-butoxy)carbonylamino]-2-thiophenecarboxylic acid 11. (i) Et3N, Boc2O, DMAP, acetone; (ii) 50% NaOH, MeOH. B) Synthesis of 3-methoxy-3-thiophenecarboxylic acid 13. (iii) K2CO3, CH3I, acetone, acetonitrile, relux; (iv) 50% NaOH, MeOH. C) Synthesis of 3-fluoro-2-thiophenecarboxylic acid 14. (v) nBuLi (2.2 equiv), THF, -78 °C, 0.5 h; (vi) (PhSO2)2NF, THF, -78 °C →RT.
137
gave 5. Crude products were purified by reversed-phase HPLC and characterized
by MALDI-TOF mass spectrometry.
Figure 5.4 Synthesis of hairpin polyamides. (i) Synthesis of polyamide resin by standard solid phase techniques;14 (ii) TFA, CH2Cl2; (iii) 3-R-thiophene-2-CO2H, HBTU, DMF, DIEA; (iv) TFA, CH2Cl2; (v) Ac2O, DMF, DIEA; (vi) Dp, 40 °C; (vii) PhSH, NaH, DMF, 100 °C.
138
DNA Binding Energetics
Quantitative DNase I footprinting titration experiments (10 mM Tris-HCl, 10
mM KCl, 10 mM MgCl2, 5 mM CaCl2, pH 7.0, 22 °C) were performed on 5’-32P end-
labeled, 285 bp PCR product from plasmid pCW15.10 This plasmid contains four
binding sites that vary at a single N-terminal position, 5’-A T N T A C A-3’, where N =
T, A, G, C. The DNA sequence specificity of novel thiophene-2-carboxamides was
evaluated by comparing their affinities for each Watson-Crick base pair to those of
N-methylimidazole (Im) and N-methylpyrrole (Py) (Figure 5.5 and Table 5.1). The
divergent behavior of control polyamides 9 and 10 illustrate the need for
development of new N-terminal residues. A terminal Im/Py pairing 9 binds its match
sequence, 5’-A T G T A C A-3’, with high affinity (Ka = 7 x 1010 M-1) while showing >
15-fold preference for G•C relative to T•A, A•T, and C•G base pairs. Terminal
Py/Py pairings 10, on the other hand, are characterized by little sequence specificity,
binding T•A, A•T, and G•C with comparable affinity.
Within the thiophene-2-carboxamide series, an unsubstituted thiophene ring
Tp 1 paired with Py shows little sequence specificity. Addition of a methyl group at
the 3-position exerts a dramatic effect on sequence specificity; A,T favored over
G,C. Polyamide 2 binds both T•A and A•T with a 140-fold preference for T,A
relative to G,C. Amino 3, acetamido 4, or hydroxyl 5 substituents at the 3-position of
thiophene all distinguish T, A from G, C but again do not distinguish T•A from A•T.
Remarkably, a 3-methoxythiophene 6 paired with Py shows good affinity for T•A (Ka
= 2 x 109 M-1) with 6-fold selectivity for T•A relative to A•T and > 200-fold specificity
139
relative to G, C. Fluoro 7 and chloro 8 substituted thiophene paired with Py afford
higher binding affinities for T•A, but a lower selectivity (3-fold) for T•A over A•T.
Figure 5.5 Quantitative DNase I footprint titration experiments for polyamides 1, 2, 6, and 8on pCW15 PCR product. Lane 1, intact DNA; lane 2, A reaction; lane 3, DNase I standard; lanes 4-14, 1 pM, 3 pM, 10 pM, 30 pM, 100 pM, 300 pM, 1 nM, 3 nM, 10 nM, 30 nM, 100 nM polyamide, respectively. The chemical structure of each N-terminal residue is included at the top of the gel and the four binding sites are labeled.
140
Molecular Modeling (Table 5.2)
Molecular modeling was performed using the Spartan Essential software
package.15 N-terminal residues were first minimized as methyl-2-carboxamides
using an AM1 model. The resulting geometry was then subjected to ab initio
calculation using the Hartree-Fock model with a 6-31G* polarization basis set. The
partial electrostatic charge of the sulfur atom, δS, and the partial charge of the
peripheral atom of the 3-substituent, δR, were examined for each novel thiophene
residue. The electronic influences of 3-substituents on the polarization of the sulfur
atom follow expected trends, with partial electronic charge, δS, decreasing as
follows: 4 > 7 > 1 > 6 > 5 > 8 > 2 > 3. The electronic surfaces presented by the 3-
substituents, δR, were also calculated and found to decrease as follows: 5 > 3 > 4 >
1 > 6 > 2 > 7 > 8.
The relative energy differences between minimized syn and anti
conformations were also examined for each new thiophene ring. Hairpins 1-5 show
141
a preference for the anti, or “sulfur down,” conformation which may be attributed to
lone pair repulsions between the sulfur atom and the carbonyl oxygen of the 2-
carboxamide moiety. This bias can be reinforced by favorable hydrogen bonding
interactions between 3-substituents and the carboxamide as in polyamides 3-5. By
contrast, polyamides 6-8 display a bias for the syn, or “sulfur up,” conformation,
possibly owing to more severe electronic clashes between the electron rich 3-
substituents and the carboxamide relative to those of the sulfur atom. Finally, the
solvent exposed surface area of each 3-substituted thiophene was compared to the
unsubstituted thiophene ring to assess the steric contribution of the 3-substituent
and surface area was found to increase in the following order: 1 < 7 < 5 < 3 < 8 < 2
< 6 < 4.
Discussion
The observed equilibrium association constants for polyamides 1-5 support
an anti conformation for the N-terminal thiophene residue. The binding preference
142
of these compounds for T•A / A•T relative to G•C / C•G might be a result of
unfavorable steric clashes between the sulfur atom and the exocyclic amino group of
guanine. The binding properties of N-terminal, 3-methylthiophene-2-carboxamide
residues also correlate well with values derived from internal contexts, where the
sulfur down conformation is stringently enforced.11
It was envisioned that polyamides 6-8 would assume an anti conformation by
sterically disfavoring contact between the bulky 3-substituents and the floor of the
minor groove. Quantitative DNase I footprinting revealed modest selectivity for T•A
relative to A•T and excellent specificity for both of these over G•C / C•G. However,
binding properties of 3-methoxy- and 3-chlorothiophene residues determined at the
N-terminus do not correlate with those derived from internal positions. Furthermore,
molecular modeling indicated that unfavorable lone pair interactions favor the sulfur
up conformation. Taken together, these results could suggest that the electron rich
methoxy and halogen groups are projected toward the minor groove (Figure 5.6).
The greater size of the methoxy group relative to the halogens might account for the
lower affinity of polyamide 6 relative to 7 and 8. The greater T•A selectivity of 6
could stem from the more complementary positive electronic surface presented to
the thymine carbonyl by the methoxy protons relative to the negatively polarized
halogen atoms.
143
N-terminal 3-methoxy (or 3-chloro) thiophene-2-carboxamide residues when
paired with Py demonstrate selectivity for T•A versus A•T. This represents an
important step toward expanding the array of DNA sequences that can be targeted
144
by minor groove-binding polyamides. Confirmation of the novel binding model with
chloro (or methoxy) responsible for minor groove shape selective recognition, not
sulfur, must await structure studies in solution by NMR.
145
Experimental
Materials
Methyl-2-chloroacrylate was obtained from Acros. Benzenethiol, di-tert-butyl